Talk at Optimization and Statistical Learning 2023, updated in 2024.
The associated paper is:
Understanding the training of infinitely deep and wide ResNets with Conditional Optimal Transport
Raphaël Barboni (ENS-PSL), Gabriel Peyré (CNRS and ENS-PSL), François-Xavier Vialard (LIGM)
https://arxiv.org/abs/2403.12887